Mentions:

Map 🐞:

on fun maps??💀💀💀💀

Okay, we can't even add them in theory

for testing: what should not be in the maze

• no inaccessible areas
• no closed loops (there should be only one way to get to a cell without backtracking)

(PLEASE send demo where it happend i cant do much without it)

optimized the map a bit

send the right name

pls

Is it normal for the maze to already be formed, but for me to still be teleporting?

This map is really fun. The only thing I didn't like was that it took 2-3 minutes. Would it be possible to reduce that time?

Ge

Ge

add dummy for parallel processing

2 tee core cpu

optimized the map a bit

the latter is not a requirement for a maze to be a maze. mazes can still have multiple walls not connected to each other, so that is up to the mapper

There are different labyrinths, we made the simplest version of the labyrinth and it shouldn't have this

There are just many variations of manifestations for the second point, for example, here is one of them

please dont use draggers, theyre really slow

if that's so, then so be it, if you want your mazes not to have that. just... some clarification wouldn't hurt, since both configurations are valid mazes

are you guys trying to do randomized depth first search ?

what would be the better alternative?

i think that is what its called

with the way the algorithm works if there are loops then something is broken

literally anything else that doesnt use draggers

but how do i check if the switch if closed without them tho???

i guess flipping it over would work but falling is not that much faster

so what youre trying to do is a simple "if x closed then..." ?

yes

also else:

theres plenty of maps to take inspiration from

idk i dont play ddnet

this is not the absolute fastest but i like doing them like this

made the maze generation faster

😢

ez

bad bot

Sowwy :c

made the maze generation faster

forgot to delete testing stuff

forgot to delete testing stuff

added missing 27tp

added missing 27tp

thx

• fixed bugs

• fixed bugs (Can you make it so that

can send too?)

???

try this pls

i think this can be better if the generator generate 5 stages for example

like 5 rooms

2 min generating for -10 secs map its bad 😭

too the color combination can be better too light

it literally doesnt even make sense.

not enough switches for that

1 tee = 2 min 2 tee = 1 min

it takes max 50 seconds what do you mean 2 minutes

1 30 seconds

can take

One person told me that the map is too dark

💀

try 30 sec max and ready

it can theoretically take the age of the universe

that the other rooms reuse those switches and that all the mazes are fine 🔥

4 clients for the best maze 🔥

On the time calculator, division takes longer to generate, an interesting fact

1 12 seconds 🔥

0 / 0 ez

uhh no

but here i need wait 1m to play 10 secs

the calculator

i dont wait to play any

i wait a number

if i want speedrun this

i need wait 1m

to have luck

of a good maze

and do nice time

ok sorry

exist the prob to have all switchs off

?

in the info and fun can add turn off entities

or min in the fun room

I don't know how you generate 2 minutes, it usually takes 50 seconds. Maybe you're just catching a super mega ultra big unluck

no

no

its 50 secs same sorry

is this entities error or why in the unnhok part all the space block are air and in the hookable some are unhook and others ai

air

insta decline

• design fix

I think it would be cool if the channel owner could also click on the blue checkmark to approve maps from other people

added key that open finish

please just make me a channel owner

🔥🔥

Add timecp on keys

That's smart, until the next update

• made both keys required
• added time cp on keys

• made the keys bigger
• added HD

good map rls

yes

ez

I do 6🗣️🗣️🗣️ (ver with no keys xdxd)

ummm, where's the release?

fr

The time it takes to generate the map, but you change seconds to days 😄

imo the title logo at map spawn looks very awkward (it reminds me of a mix of mspaint drawings and warped AI-gen text) and should prolly be redesigned in vector graphics. it's just very inconsistent with itself and not in a way that feels particularly intentional.

the other text should be more legible, dark blue + black on a dark background is very hard to see

I do really like the concept for the rest of the map's visual design though, and the use of the dragon curve is a very nice touch (though again it kinda disappears if you actually zoom out to see it)

I can just delete the map logo if it doesn't fit

made the text lighter

you've made a mistake letting me see this channel

ill post all 17 optimizations soon (watch me be right and there's actually 17)

^now that's gonna be the fastest possible design

ye this is very not efficient xd

not sure where to start

doors are assembled vertically and also checked in that direction

on the phone rn but i start cooking soon

any particular reason why these are 3x3 and not 2x2? just disable gravity and the corner will work for 255.

its odd, some places you use tune but not always

questions about maze to him

ok sorry xD

np

cammodudes idea

struck by cammodude's designs again

a second cammodude design has struck the towers

whats the first one

you can more efficiently check these (and compact this) to 1 tile per switch vertically

they all teleport to the same place though so im not sure why these are different switches yet, i think i need more time to understand how this works

it checks if any of the neighboring cells arent visited

makes sense

so you basically have a 4-way OR

yes

most efficient way would be this, pretty sure. use velramp_curvature 1 and gravity 32 (1 tick per tile movement downwards)

tp in at the top, check right, left, right, left (reason for this is that doors affect X speed for more than the collided tick)

also you probably need air_friction 1 so that there's no friction for the tee

vel_ramp

for all but the first tp in, you need gravity 0 to not add extra downspeed per tick

if you prefer it's all the same tune, i can find a tile that works with just speeder

i dont think i care about that

there's also a minor quirk with your design (not that it matters i think?) which is that lower switches that are off cause the higher ones to only block 3 tele

because of door overrides

so depending on the state of the bitfield you'll leave in a variable number of ticks

okay

you have the lowest blocking 4 tele, so assuming it always exits in 3 it might be okay

45 speed with 45 angle works for small distances. im always in the highlighted tiles

works good for me. tune is grav 0 velramp curve 1 air friction 1

45,45 32,180 32,0 32,180

you can put door length in walls (TIL)

this is both a space and time improvement

working on 4x2 design for these 3 switch triggers

you can also save some time with blue tele (you can keep downward speed on the tee the whole time then you never have to waste a tick speeding up)

3 switch trigger. angle is still 45, but uses 136 to go in a 3x3 (either hits the door and goes down 3 triggering the switches or goes 1x1 down-right and leaves

ive done 2x2 myself i just wanted it to be more understandable as a screenshot

this is a replacement for this

troll

if you prefer a cleaner design for the first, you dont need all the tele. makes it easier to see the doors

conveniently i already have this 2x2 design from a mafia/among us style map im working on, its so small you dont even need tune but you might as well put it in there to be cconsistent

ignore the switch in the tpto its for that design

oh i see you disable 226 to toggle between the states?

i guess then just delete the 1 off here and keep your 226 where my 255 off is

the only remaining thing i dont understand is tp 82. are you checking all of these to make sure the maze is sufficiently completable?

if all the cells are visited

can you shortcut to teleport back in the middle, or it always has to re-eval from the top

i guess

had this idea but i think it wouldnt be that much faster

☑️☑️☑️☑️☑️☑️☑️☑️

is this one per cell? if you're always turning it off cant you do binary

or wait, always on

i need to restart the map and see what happens xd

always on

so basically you're checking each door sequentially but it starts at the top for the 226 check

you could use a binary adder for this which would turn 81 switches into like, what, 6? xD

or 7 im not sure

SIIIX SEEEEVENNNNNNN

https://tenor.com/view/sixseven-six-seven-six-seve-67-gif-14143337669032958349

but it doesnt go from top to bottom

yes 7 switches. so i can show how to make a 7-adder

its random

so for this, they can be randomly on?

dont they all have to be on to reach the bottom?

yes

they're just turned on in a random order i guess

yeah ok that makes more sense

they represent visited cells

ye

but like, they get the same tp no? so adding 1 to an adder is essentially the same as making any random door closed

assuming you're making sure you're not closing a closed door?

its for the other checks

too

like is it guaranteed to turn one on?

ah

ok one sec ill keep looking

i guess i could just make other checks check if the cell has any of its doors open but thats kinda stupid

and probably way slower

yeah i see

i love a good puzzle :3

if we're doing math right (this is unlikely im tired) worst-case for the 81-checker is 3.888 seconds and you enter like 81 times, with an average of 1.944 (curly is my numbers person i just do the talking)

this doesnt include 226 check but meh i forgot

Hi gogp

hello ramaa

with 226 check it only falls ~7 times (i made the number up)

are you not double a now?

oh i see

is it rama now

i just realized this gif doesnt just repeat after the first lick cus there's some frames u see thru the tongue

brb i gotta pee

Is Ramaa but I miss the change name and in a random server I change full profile and not sv xddddd

And I lose the wings

Of my old name xdxd

huh, i am still really confused but i will say ramaa again

it might be kinder to those with bad computers to design the outside with stoppers instead of doors, since they never turn off

i dislike the stoppers here since they're not needed like the rest of the maze, you should be able to reach here with 9 length doors?

and if you're trying to have a diagonal angle from the left or above ur probably gonna hit these

i think actually, much more than the 81-tick dropper, blue tele is going to save you so much time since you tp so much

you basically make the assumption that at each exit you're going to restore some default vector for checks, probably 1x1 down-right

then you dont have to waste the entry tick getting the speed again

you could actually use 3x3 down-right always since you just have extra speed for 2x2s

what

sorry, this speed. 136 but same angle (45)

so you can always use 3x3 speed in your tele exits (assuming you also stop the tee on that tick, with solid tiles)

basically, speeders are evaluated in the same tick that you would teleport. so if you put a speeder in a blue FROM tile, they will be teleported to the tpto translated by that speed on the first tick

tele actually has a very low priority. that's why putting switches like this (https://discord.com/channels/252358080522747904/1477444865410203678/1478453120810876988) below where the tee should be able to go even works, since switches are checked before teleport is handled

okay i probably need to sleep to understand

it's okay, and i may make some edits to the map when i finish the 81-tick drop thing

i do want to give a compliment that this map is really well organized, it seems like a lot of thought was put into the tele/switch allocation

and ofc that this even works is amazing

another thing we realized is that these two steps are mapped one-to-one which means the teleport step is not really needed

so you could instead of jumping from a 37 to a 227 (which is always done if the door wasnt on) just put whatever is at 227 below it

this saves 1 tick for every other tp not spent in the drop check

any reason the top right and right edges are not connected like the rest of the cells? it seems like you only chose to connect the top-left and top row edges, but it would look cleaner if they were all connected like this

ah and you did connect the bottom row and bottom-right

maybe you did it for alignment? but it would still be aligned since this top row is offset by 2 tiles.

(yes im working on pointless things cus i hit a dead end but dont complain)

if the top doors are always on, why check them in the mech?

just start checking from 1,2,3,etc.

so corners only need to check 2 door states, and edges only need to check 3. so these can be 50% or 75% height respectively

partially done

with the exception of the completed check, tees always have some kind of speed before they teleport. in this case its [32,32]

we also start checking neighbors immediately. for each cell we check the left cell, or right cell instead if we're on the left edge

the completed check is inverted now. we can't move down in the same tick if we would be stopped by a door moving downwards in here, so we decided to just make an elevator instead :D

note that 226 is here along with its 81 groupies (9x9 squad)

oops i forgot the speeders so that looks confusing probably

note that if we move left, we count on a rightwards force in the rest of the mech, so we have to fix this by applying 64 right.

gravity is inverted for direct teleports to 82. ill explain why we cant preserve speed for some 82 teleports later

might be the grossest thing ive ever made in ddnet? but here we are 4 hours later i guess

for 1-9 tpto, gravity is doubled

(which lets us set 3 switches in a tick still)

i was very confused for a while until i realized there's an entities bug

seems these switches weren't copied correctly, left half of the last 3 columns should be the same as the right half

i believe this causes this boundary to always be open

er, if 11 is ever directly accessed it always turns off (the cell below)

i rq maybe i try more tomorrow

the leftmost is for upper cell center left for left center right for right and rightmost for down

i just deleted ones that are not nedded

what even

i was too lazy to do it correctly

so last night i had a revelation while trying to fall asleep, still playing ddnet sudoku in my head, and i think its best if i stop trying to space-optimize this

which also means the map can be readied faster

alright, sorry for not working on it yesterday. ill try to get this done

it will also not look like this ^ i think i made it too messy

damn wtf i just got this rigged ass maze

its like comically bad

Why no a little change if you don't touch a button add you 1 minute and u can finish without keys🔥

i think its always better to get the buttons then, it doesnt take that long

probably friction 1 of either kind and velramp doesnt belong in the maze itself

otherwise its possible to get through the doors with speed

ill fix this in my version. was it just gravity 0 before?

minor entities bug

happens in the whole column

but doesnt that just affect x?

whats the point of having different x and y physics in 0g?

for the mech you do want same physics

but not really for the maze, i think its best if its only gravity 0

if you want no friction as well that's fine, but uncapping velramp in maze lets you get speeds you shouldnt get

in x direction only, yes

i guess you can still skip through doors vertically, which might help you get to the buttons quicker

i guess if you are comfortable with potentially more door skips, then it doesnt need to be different tune

it just seems pointless to have slower speed only in 1 axis

it would be better to somehow make it in both but probably not possible

yeah you cant enforce a velramp-like thing for Y

so ye only thing you can really do is just disable and maybe cap max speed u can get with jetpack

it is a shame there's no friction for Y

found more bugs

7 is checked as the cell above for two cells side-by-side

second one should be 8

where's 54 here? 72 is here twice

can this be moved to waiting? both for bugfixes and if i choose to finish optimizing xd

releasable as it is if the bugs are fixed but i would like to optimize it since i think it can be approx 5 times faster maybe more

maybe

maybe.

okay

done

i mean i posted map bugs and no mapper response in like 2 weeks so

xD

alright thanks

we thought you were working

fixed 3 incorrect switches and an entities bug

i hate this stupid bot

He accepted it quickly this time, don't offend him😠

bad stupid idiot bot

Why didn't it exit the waiting mapper?

because bad bot

:(

• fixed a thing

• fixed a door

pog

yea maybe give it like a week and if im too lazy to upload fast version then just rls

i feel bad for making u wait for nothing

ready?

how about test and dont be funmapphobic :(

i mean this is like the one thing u said and i was gonna fix xd

I think this good idea for fun map

thanks :3

i still think it could really use a better logo if that hasnt been changed still i could take a shot at making one but it might be a bit cause im busy

he just googled maze font and added gradient and outline btw

If I had drawn it myself it would have been much worse💀💀💀

still working?

i havent worked on much since, been busy

i do have this semi-optimized version that just changes the 255 picker

ideally all of the tele to single destination has blue tele

hm, my version is buggy D:

oh i know what i messed up

• converted 255 to blue tele

most of the mech happens in the neighbor checkers so its not a big improvement

it is still possible to die

yeah im really confused. i thought it was just a hot_reload bug

also it seems like it can take a while to find a free cell when it goes back to 82

dunno if this is also worst-case for original map

i guess im just bad!

i think these doors being uneven spacing in the tall section and in the neighbor checks is becoming a problem

ill change the tall section to zigzag which should fix some of this

everything is fine, don't be upset

i also mentioned to curly we should check for completeness in a non-linear way. so we should check over large hops to see if tightly-grouped sections are probably done

so instead of checking from 1-81, we should check maybe 10 first (4 neighbors) then opposite corner from that, doing big hops throughout the cells

• silly broken 82 check (hits top of doors that are lower, need to make it taller probably)

i can look again after work but the idea is there

i can approve?

Can I approve?*

nah your not tester

And you an ignorant person who doesnt know how to speak english

It's a joke, relax man, I know I wrote it wrong with my intermediate English 🥀

I can see the asterisk at the end of the sentence.

Translator final boss

wdym approve i said broken

it just feels like u would be approving a map u dont understand, leave the task to a tester that can tell if the map is ready please.

its also not my map, youd have to ask the author

bro i have proof in my dms from you that u are not one to talk 🥀

tbh same thing happens with volley. i ask people to test it and just cus its a Fun map no one thinks they can actually test and find bugs

but ppl find entity bugs in much larger maps and the things i found that were wrong were not harder than that

who tf accepted a broken verison???

Fr🔥

its like sending it on an unfinished college presentation that doesnt even load

Should we just wait? I just don't quite understand what's happening with the map

i said i dont really have time (honestly energy) to keep trying to optimize it, if you are both happy with releasing it unoptimized i think the map works if those things were fixed

i had a random spike of energy there but it kinda died

i think retsot should swap it back to letting you get same speed for X and Y in the maze, it let you skip more places but it did feel nice

i know i said to not allow it but you cant prevent them from skipping doors going up and down

if both of you are happy with it, i will test a final time and just make sure there's no bad bugs

sorry to be confusing and annoying and insert myself into the process ><

im probably going to optimize it but idk

im struggling, the tee comes from 85 with speed of 1block/tick downwards, if the bottom door is active the tee looses downward speed, can i somehow make the door only 1 tile or make it work somehow else? (tune removed for readability)

ok, making it 2blocks/tick solved it but its 2 times bigger cuz of this

• Decreased generation time massively -# * Went insane

• forgor to remove test thing

• forgor to fix fun and info tp

• fixed a bug that would cause you to get stuck during generation

i didn't find any more bugs, the generation time was reduced significantly, there were no complaints about anything else, so i think the map is ready for release

• further reduced generation time

• changed finish slightly

• changed finish slightly (again)
• fixed some minor entity/design bugs

the way that blue tele works is that their teleported position is translated by the speed they had when they entered the tele. so, if you have a unit speed of +1,+1 then you will spawn inside that door.

you can make 1 tile doors by putting the Extend tile into a solid, if that's a separate thing you want

sorry for the late reply, ive been taking a break from ddnet

TLDR? or is it too much and i should just go look

i assume you already know about doors not allowing you to move in a unit direction (-X, +X, -Y, +Y) for an additional tick after collisions

so you have to be careful with that when doing high-speed mech, otherwise you will assume you can keep pushing them in the same direction. that's why the mech in volleyball currently is so complex, we have to keep choosing a new direction to move them in for it to be fast

well, theres the cube now

my problem was that the tee would get stopped if the 229 door was active, the solid tile would just always block them

actually it wouldnt but now its only possible to activate/deactivate 1 switch and i need 2

• made the mech considerably smaller

uhhhhhhh wow you really changed it a ton

i think i can assume just by looking at this that it is in fact optimized

like the way you mapped it seems like you have total understanding of the best ways now

assuming no bugs found in next 2 weeks, this should be marked for release IMO

great work!

even adjusted the ending so you dont hit stoppers

• so you have 254 and 255 so you can deal with the cases where you're on the right and left separately? i guess that's one way to handle that
• when i was redesigning the "done" check i scrambled the order so that you check the farthest cells from other cells you check in sequence. i notice you still start at 1, after giving them 2t speed down. if you want to further optimize it, i think you should start the tee in the center of the grid, then do visited checks with maximum total hop distance

basically, you come up with some sequence where you always pick a new cell based on how many neighbors are "probably" not checked (this is just based on the first TP in, which i think is your CP in this case)

i can post the order i came up with, it wasnt perfect though and you might be able to bruteforce or algorithmically define some better order

my list also skipped ever checking the first cell, since you always visit it from spawn

order is based on breadth-first search, and i just crossed off numbers in a drawing program, so it might not even be perfect (this assumes the player has already visited the middle cell, 41)

11, 71, 65, 17, 29, 53, 13, 69, 15, 67, 35, 47, 25, 57, 61, 21, 43, 39, 23, 59, 31, 51, 49, 33, 75, 7, 79, 3, 63, 19, 27, 55, 5, 77, 37, 45, 1, 81, 9, 73, 8, 74, 6, 76, 4, 78, 2, 80, 10, 72, 28, 54, 46, 36, 64, 18, 66, 16, 68, 14, 70, 12, 62, 20, 44, 38, 26, 56, 24, 58, 22, 60, 30, 52, 48, 42, 40, 32, 50, 34, 41

i think even if you just shuffled 1-81 it would be better than checking in order xd

if you do decide to do this btw i dont think u have to change the cp/cto, just the random neighbor tps and the switches

does it matter? it checks every cell only 1 time either way

i might make it start from middle cuz it seems to make more interesting mazes tho

might have found a bug actually

i forgot tune in one of the corners

if you try to check cells that probably have no neighbors checked, it will fail more often, taking you back into the mech sooner

you want your time in the dropper to be low

and only when its probably complete

so at first i was checking in a knights-move pattern, then i switched to a pattern that jumps between corners counter-clockwise

and i found that generally it should be done before it even passes through 6 of them

but it always checks 81 cells and sends to the mech the same number of times

the final pass it will always check yeah

it can stay in the mech longer if it starts in the center

since it can branch more

im confused how is starting position related to the dropper check pattern

im saying that the pattern doesnt matter

i might not understand correctly but i thought it can loop through until it finds spot where all the neighbors are already visited

like if it folds onto itself it goes to check if its done, and then it gets a randomized position again

the done check only checks a cell if it wasnt previously active when doing the check thats why all the checkpoints are there

ye so its possible i just dont understand it right

so it always goes through in constant time (gets next CP, eventually through to bottom)

yeah

• bug fix

before i disappear again i think this map is great and just needs more people to test, then release fast

retsot literally implemented all of my feedback and made it fast so its definitely ready

Dont disappear :)

hi please release this map its great :D

hi please release this map its great :D

hi please release this map its great 😄

it seems there's an extra setting here, but i think it has no impact (maybe it will be auto optimized when releasing?)

• ^

now release for sure

nixus was gonna ready this like 3 months ago and now that it's had the biggest glow up a map has ever seen it's like it doesn't exist

like, ready to full send, unoptimized untested

now that it's polished no1 cares xd

theyre waiting for deco overhaul

change everything

even the concept

then release

deco overhaul?

Maybe the testers just aren't testing the map itself, so they can give it a ready

complete redesign (not happening btw)

movement is a bit hard. can you make so the tee doesnt get so much momentum??

wdym? reduce the jetpack strength?

or air resistance

i dont know how it works, but you shoot a bit and you fly with no friction like it was space

turn off entites

?

the design is literally in space

yes but theres 0 friction, its just unconfy

if at least you put a bit of friction is ok

i kinda want it to be identical in x/y directions and theres no friction for y so idk

yeah i think lowering the jetpack_strength would be nice since you kept the velramp stuff in

would still play naturally x/y like you say

both for making it easier and not easy to skip, 2 birds one stone xd

• set the jetpack strength to 200

tbh i think it makes it even less controllable but i wanna hear what others think

ill test it

definitely feels too slow now, if there was a way to set a max speed that would be cool, but yeah maybe try 300

i still think a max speed tile would be a great addition to the game, since speeders always impose it in some direction and the force cannot be disabled

• 300 jetpack strength

better to add unsolo for more fun. The laser isn't make sense because there is no freeze in this section

make this half hookable to save some time

endless hook

just kill yourself (in game)

hi please release this map its great :D

if no one can manage to get stuck with 300 its rls for me

if you have to travel like 6 cells or something in a row (really lucky) in order to clip i think its fine

you can skip only if all except the last door are open

like this

yeah that's so like, doesnt matter at all! rls!

cool map

rls when

?

Fun maps can release at any time right?

ye

really. its a cool map

you can skip with less, also works horizontally, would be nice if you find a way to fix it

if you can't fix it then I'll ready it as it is now

i dont think there is a way to fix it tbh maybe someone could make better door and stopper tiles in the future so that stuff like this doesnt happen

i think i was accidentally testing with jetpack strength 200

such a cool concept

fun to play

https://cdn.discordapp.com/attachments/1257742065056485458/1533917812777553951/giphy.gif

https://tenor.com/view/oneshot-oneshot-niko-niko-kine-master-gif-23583473

gg guys

hi, could the map be released on September 19th at 09:09 UTC?

9999999

and 09 seconds

🥰

there is a tournament on the 20th so idk i guess

needs to decide 😅

its a fun map

i mean its fun so it doesnt rly matter, if u dont mind a fun release the same day as your map

(there wont be any announcement for the fun map tho)

ahh yeah i forgot

just wanted to remind everyone of the tournament before the decision is made

actually i realised that my map also releases on the 19th

yeah whatever 😅

ye, also if u want we can move ur map to avoid getting overlapped by tournament (didn't know about it either until it was announced, sorry for that xd)

i also didnt pay too much attention.. btw technically you have access to the channel where we planned the tournament date, but yeah i didnt think about tagging you either.. i guess moving my map to the next free date makes sense, but i am fine either way

ye thats my fault i'll pay more attention to these channels

next free date would ideally be <t:1791637200:F>. does that work?

either way I scheduled your map for it, there will be ingame broadcast and should be discord message if bot doesnt bug

Sounds good, maybe 3-4 hours earlier would be nice so the chinese community has more time to play the map before hopping into bed, but again i am fine either way

ok ill make it 11 AM our time

I'll edit the current message in

, i don't like that it's not in release order but deleting and posting again is even more confusing i guess lol

https://cdn.discordapp.com/attachments/803276311182573608/1479359764562448454/man-break-chain.gif